A scan told me I have a bad throttle sensor. Could it just be a loose wire and how can I find it?

The fault didn't say the sensor was bad. It said the voltage was too high or too low. Sensors are only responsible for setting codes about half the time. You're one of the few people who are right to suspect wiring problems first. Measure the voltages on the sensor wires by back-probing through the back of the connector while it's still plugged in, then holler back with those numbers.

I can't find a decent picture of the sensor but it appears to be part of a larger black plastic housing on the side of the throttle body assembly. There should be three wires going to it.

Ground the black meter lead to the battery negative post or a paint-free spot on the engine. Poke the red probe through the rubber weather seal on the connector. You can use a stretched-out paper clip too if necessary. What you should find is one wire reads about 0.2 volts, one reads 5.0 volts, and the middle one will read from approximately 0.5 at idle to 4.5 volts at wide-open-throttle.

If that middle signal wire always reads close to 5 volts, the ground wire is open, (broken). If it always reads near 0 volts, the 5.0 volt feed wire is open. The signal wire is a little trickier. If it is open, the voltage swing will be correct at the sensor but at the computer it is designed to be "pulled up" to 5 volts or "pulled down" to 0 volts to force the computer to detect the problem and set a code.

The easiest way to check the sensor's operation is with a scanner that displays live data. You can simply sit in the car and work the gas pedal while watching the TPS voltage. If it swings normally from 0.5 to 4.5 volts, all of the wires and sensor are okay.

I just looked up the exact fault code description:

P1122 †Throttle Position (TP) Sensor Inconsistent With MAF Sensor Low Voltage

That doesn't indicate a defective sensor, ... Exactly. It means the two sensors don't agree on some engine operating condition. The MAF sensor reports incoming air by weight which is low during idle. The throttle has to be closed and the TPS reading should be near 0.5 volts. The TPS is a relatively simple mechanical device that for the most part can develop an intermittent connection between its movable contact and the carbon strip it rides on, but the MAF sensor has a bunch of electronic circuitry inside so it is more likely to develop a problem like what the fault code is describing. Either sensor can cause an intermittent problem too. The fault code should always be recorded, as you did, then for intermittent problems, it should be erased after any repairs to see if it comes back. Codes will be set by unplugging any sensor while the ignition switch is in the "run" position too so do any work in checking connector pins with the switch off.

Given the code and the history of events, you might look closer at the tube between the mass air flow sensor and the throttle body. Any gap in that tube that lets fresh air in will cause a problem. The throttle blade will be open allowing more air in, but if any air goes in that doesn't go through the MAF, the computer won't know about it and won't command sufficient fuel to go with that air. Based on the throttle position sensor's voltage reading, the computer expects to see a corresponding air flow reading from the mass air flow sensor. THAT is the discrepancy the fault code is referring to.

Absolutely, but only if it is after the mass air flow sensor. With a hole that big, it's surprising the engine would get any gas at all and run.

There can't be any leak between the MAF sensor and the throttle body.

As for the oxygen sensor, it only detects unburned oxygen and if there's too much of that, it's going to instruct the computer to alter fuel metering. By unplugging it, while it will change how the engine runs, it's simply eliminating some of the information the computer uses to base those calculations. Those sensors don't even work at all until they get up to 600 degrees, so if the sensor was the problem, the engine would run fine for the first minute or two.

"Air to fuel ratio sensor" is just another name for oxygen sensor. No sensor actually responds to unburned fuel in the exhaust, just unburned oxygen.

Not sure what other code you're referring to but unplugging any sensor while the ignition switch is in the "run" position will set a code.

Normally when a sensor is known to the computer as being incorrect, as in unplugged or anything else that sets a code, it "injects" an approximate value from memory to try to run on as best as possible. It bases its fuel calculation on other known factors like the mass air flow, engine speed, throttle position, and things like that. Once it's warmed up, the amount of fuel being called for is tweaked a little based on the oxygen sensor's readings, but the computer can only adjust the amount of fuel up or down about ten percent beyond the factory-programmed values. That is not enough to cause running problems on its own.

What can happen is the computer can try to overcome what the oxygen sensor is reporting. This is where the diagnosis gets complicated, and even experienced mechanics get confused. Suppose, for example, you have a defective spark plug or plug wire that causes a misfire. Unburned oxygen and fuel goes into the exhaust but only the oxygen is detected as a lean condition. (On '96 and newer models the computers can tell which cylinder is misfiring and set the appropriate code, but lets ignore that for now). Normally all the computer knows is there is too much oxygen in the exhaust but it doesn't know why. In response it adds fuel to all the cylinders. No matter how much extra fuel it adds, there will still always be that unburned oxygen from the misfiring cylinder. Eventually the computer will figure out something is wrong and set the code "running too lean too long", but you smell unburned gas at the tail pipe.

The computer is constantly switching from too rich to too lean a couple of times per second, partly to make the catalytic converter work and partly to make sure it has control over the fuel / air mixture. It's that constant extra air from the misfire that makes it think the system is running too lean and it has lost control.

It is possible that by unplugging the oxygen sensor, the computer knows it can't adjust the fuel mixture based on those readings so it doesn't even try. It's not really the lack of the sensor that makes the engine run better. It's the lack of information being reported so the computer doesn't react to whatever is really causing the problem.

Think of someone pretending they're going to hit you. You react by flinching. But if you cover your eyes, you won't flinch, therefore covering your eyes means they never started swinging their fist at you. See the flaw in that logic? The cause is still there but you blocked the signal so you didn't react. With the computer, the cause is still there but the computer may be not correcting for it because it isn't aware of the RESULTS of the problem when the O2 sensor is unplugged.

Can't argue with success but it would be nice to know what the problem was. Will keep my fingers crossed.

Could that have allowed some air to sneak in?